The present invention relates to methods and systems for distributing information between a plurality of nodes or connections using an optical fiber, the fiber preferably comprising a core, a cladding and a protective coating, the core and cladding preferably being made of glass, the coating being a buffer made of a polymeric material, the fiber being either single or multimode, and step or graded index, the invention being applicable to both glass-on-glass fibers and plastic clad fibers.
Prior art methods for distributing information between a plurality of nodes or connections have generally relied upon the use of copper wire for creating a plurality of point-to-point links between a central node and the plurality of nodes to which distribution is desired. For example, for telephone service, the central node generally corresponds to a telephone central office, a private branch exchange (PBX), or a feeder station.
Since optical fiber has extraordinary high band-width potential, numerous concepts have been proposed whereby a single optical fiber is utilized by a plurality of nodes for distributing information therebetween, the advantage of such concepts being the use of a shorter length of communication cable as compared to architectures utilizing point-to-point links, a further advantage being that the cost of components in systems embodying such concepts which are used by more than one node is shared by all the nodes using that component so that the per node cost of such equipment is correspondingly reduced.
Notwithstanding these advantages, actual distributive optical fiber communication systems have remained largely a laboratory curiosity due to various technical problems in implementing such concepts using optical fiber. One problem has been that of tapping optical fibers so that each of the plurality of nodes can withdraw information from the optical fiber and inject information into the optical fiber in an inexpensive and yet efficient manner so as to render the system capable of serving a sufficiently large number of nodes to make the cost of the system competitive with electrical systems.
To date, various tap proposals, in addition to being unduly expensive, are also problematic when long term system life is desired since it is known that static fatigue is detrimental to the integrity of optical fiber. In fact, the problem of static fatigue failure of optical fibers has long been recognized by the prior art and has long been a critical design criteria evaluated when implementing any kind of static optical fiber telecommunications system. For systems wherein the fiber is to remain in a static environment over a long period of time, the prior art teaches that the maximum long-term stress the fiber is to be subjected to be of the order 0.2.sigma. to 1/3.sigma., .sigma. being the proof stress of the fiber shortly after being manufactured, the proof stress for typical optical fibers generally being of the order of 500 kpsi to 1000 kpsi, such proof stresses generally corresponding to 0.5% and 1.0% strain respectively. Accordingly, the prior art teaches that for long term static optical fiber telecommunication installations, the maximum strain any part of the fiber is to be subjected to be of the order of 0.33% strain and below (see page 94 of Conference on Optical Fiber Communications, Mini-tutorial Sessions '86; and Skutnik et al. "High Strain, Reliable, Hard Clad Silica (HCS).TM. Fibers", presented at FOC/LAN '85, page 235, left column lines 1-2).